Why You Hear What You Hear


Home | MAX patches | Software | Using Falstad applets | Recording Audio| Links

Supplements for Chapter 7:

Sources of sound



The Moodus Noises are back! Story here. Jelle Zeilinga DeBoer, a professor emeritus at Wesleyan University, studied the Moodus Noises for years; he wrote a book, "Stories in Stone: How Geology Influenced Connecticut History and Culture."

Cookie cutter phenomenon here, and the YouTube video of the newscast.


Sound files:

From Why You Hear What You Hear....

Sun Chips 95 dB The original compostable Sun Chips bag was recorded by the author before it went off market. You might want to analyze this sound file in several ways, including making a power spectrum, a sonogram, and direct inspection of the waveform. The microphone was about two feet away as the bag was crushed; some echoing from nearby walls may be present (you can check for that by computing an autocorrelation in, say, Audacity). The drop-off around 16 kHz you will discover in the power spectrum is due to recording limitations, not the bag itself.

Remember to adjust "size" (the number of points used in the Fourier transform, a selection under Analyze...Plot Spectrum, with Algorithm: Standard Autocorrelation selected for the Autocorrelation) to get a right range of time showing for the autocorrelation.

We applaud the Frito-Lay company for making a green, compostable chips bag, and hope that a satisfactory replacement can be found.

A hint about what helped cause this bag to be so loud, is found in this article by Maggie Koerth-Baker here, partly reproduced below:

"Frito-Lay will be retiring the compostable Sun Chips bag, that became famous less for its Earth-friendliness than for the 95-decibel crunching sounds it made whenever it was touched. Created from a polymer material based on corn starch, the bags were cursed with a high glass transition temperature. Basically, all polymers have a rubbery state and a stiff state, and each type of polymer switches from floppy to crunchy at a different temperature. For the Sun Chips bags, that was, unfortunately, around normal room temp—so what was supposed to be flexible was constantly turning brittle. And loud. The weird thing: this flaw isn't inherent to compostable chip bags. Boulder Canyon potato chips makes their version from a wood pulp polymer, that seems to avoid the problem."

Crinkling is always catastrophic: one instant you have a surface under stress but not moving, and the next it has failed and moved abruptly - requiring large accelerations. The stiff, brittle surface sustains a lot of force before it fails, and it is so light that a lot of stored energy ( force times distance the surface moves) is released rapidly to a very light piece of the surface (the material was indeed VERY lightweight) - again, resulting in huge accelerations. The surface of the bag near the crinkle moves a great distance, perhaps half a centimeter, in under a millisecond - huge compared to the mere micron displacements for a loud surface vibrating at say 1000 Hz. A sudden pulse, perhaps even a little shockwave, is created and propagates outward toward your ears. This whole failure - acceleration scenario is happening perhaps 100 times per second somewhere on the bag when you crush it. Ouch! It would be nice to get a schlieren image of the bag being crushed. Unfortunately it may be hard to obtain the material now.

Project: Sound from crinkling (advanced)

Try to determine what is happening when a surface crinkles suddenly. What is the nature and cause of the sound, as a function of crinkle size, crinkle acceleration (determined by high speed video?). How does the sound fall off near the crinkle; how much of the sound near vs. far away is of a wavelength comparable or shorter than the size of the crinkle?

Dipole near a wall

Below at the left we see a dipole radiating fairly weakly, (due to acoustical short-circuiting of the two monopoles); in the it is near a wall in one orientation, and at the right it is at another orientation with respect to the wall. There is less sound radiation in the middle, and more at the right, than the case with no wall at the left. Can you explain why?


Tuning Fork radiation pattern

Sound from a tuning fork (Figure 7.29 in Why You Hear What You Hear). Dan Russell has nicely illustrated several related scenarios involving sound radiation from a oscillating dipole or a pair oscillating dipoles (a tuning fork in the right geometry: dipoles close together, oscillating out of phase).

Project: Horn simulation, self-amplification

In chapter 7 we considered horns driven by a source at their throat, and the issues of self-amplification and confinement of the source. Beyond that, the horn has to lead the sound out with as little reflection as possible. It is easy to draw various horns in Harvard Ripple. To get started you can download this page to load the Java applet on your computer..

Using the four detectors, and saving their data, see if you can show that more total sound is coming out of the source if it is in the horn than if the source is in free space.


Sound from a baffled loudspeaker

Prof. Daniel A. Russell of Penn State University has made very compelling moving illustrations of the sound pressure coming out of a baffled loudspeaker as a function of the frequency of the speaker. He also includes measurements on a working speaker.

Music Box Mechanism

The comb and braile of a music box, on a small frame with no sounding box, is first played in air and next played with the frame pressed against a table; listen. Why is the vibration, starting at the tines, so much louder when the frame is attached to the desk? The various factors involve transmission of the vibrations to the desk, short circuiting of sound from a small dipole source, launching sound from a vibrating surface (that may or may not involve short circuiting), and questions of the listener being in the "near field" of the desk, where the sound may be louder. There is no question though that overall the desk is much more effectively coupled to air vibrations (sound) at the frequencies that are being transmitted than are the tines of the music box.

The little oscillating teeth of the comb in a music box are the ultimate source of the sound, but they produce almost no sound directly. The vibrating teeth do a good job of pushing a little air directly in front of them and pulling at the air behind, but these two regions are very close together and communicate with each other at the speed of sound, 343 m/sec. Only about 3 millionths of a second is needed for the high pressure cell in front of the tooth to short circuit the low pressure at back of a tooth a millimeter away, and vise versa. Suppose we connect such a vibrating tooth to a tabletop, allowing it to transmit some of its vibrational energy to the table surface. This surface presents a much larger area to the air as it vibrates, is partially baffled dues to its size, and does a vastly better job creating sound from vibrational energy. We thus expect the sound to become much louder if the music box is touched to a table. Music box comb abd braille are inexpensively avaiable at some toy or novelty stores, and on the internet.


Software Siren: MAX Siren by Jean-Francois Charles

Charle's MAX Siren is a flexible tool. It requires that you download the free MAX OSX or Windows runtime software, at http://cycling74.com/downloads/ Then download Siren from the MAX Patches page (link above).


Project: Sources that short circuit themselves

A simple project can quantify the short circuiting effect that is discussed so many times in Why You Hear What You Hear. The key is to show that the sound power near the source is larger than would be expected judging by the sound power more than a wavelength away..

This can be done by going more than a wavelength away from the source to make recordings with a microphone, then recording closer to the source and comparing (if the source emits many different frequencies you can filter out different ranges of wavelengths later with your sound processing software). But there is a crucial aspect of this type of measurement that cannot be ignored: a short curcuiting source will emit sound strongly dependent on angles defining the orientation of the object doing the sound production. What we are talking about here is illustrated in the sound intensity near a tuning fork, as illustrated by a Ripple simulation in Why You Hear What You Hear (chapter 7) and shown here. So sound intensity is VERY loud in some pockets near the source, but it vanishes along curved nodal lines, and it drops off in intensity rather sharply away from the forks. There is a lot of sound trapped near the forks and not escaping, but it is a job perhaps for an engineer or ambitious amateur to quantify this through measurements.


Project: Defeating short circuiting

How do you get more sound energy out of a source that is short circuiting itself? Why You Hear What You Hear discusses several scenarios involving baffles, and what we termed Near Field Capture (and release), in chapter 20.

One scenario is spectacular if you can find or make a simple horn, and buy one of those greeting cards that makes some sounds when opened. Carefully remove the little speaker glued to the card, keeping the wires attached. (The battery may only last 10 or 20 plays so be conservative). The speaker is pretty weak and squeaky. Note that it is open in the back and is therefore a dipole source. If the throat of the horn fits snugly around the front of the speaker, it will be MUCH louder than before, in free air. This effect is key to everything from an Edison phonograph to modern loudspeaker systems costing tens of thousands of dollars.

The main effect here is self amplification of the source in the throat of the horn due to confining and compounding the amplitude there by constructive addition. The sound coming out the back of the little loudspeaker is no match for what comes out the mouth of the horn, and what short circuiting is left is no longer important: you can't do too much damage to an amplitude of strength 10 with an out of phase amplitude of strength 1 (well, you can take it from 10 to 9). Also, the sound is emitted farther away, from the backside of the speaker as opposed to the mouth of the tube, another factor in reducing the short circuiting.

A simple cardboard baffle, the stiffer and larger the better, with a well fitting hole cut in the center for the speaker, will also stop some of the short circuiting and make the speaker noticeably louder. Muffling the sound from the back of the speaker will also make the soind louder, again by reducing the short circuiting.


Most people take the shape of a bull horn for granted, or think perhaps it is shaped to direct the sound where the horn is pointed. This indeed is one of its functions, but not the primary one, perhaps # 3 of 3 primary functions.



If there is a sound source at the throat of the horn, the horn's job is (1) to amplify that source by confining and compounding the amplitude there by constructive addition (also called "loading the source"), and (2) to seamlessly let the sound out (without strong impedance mismatches and attending internal reflections, that if present would cause resonances, coloring the sound). Obviously the mouth of the horn causes an impedance mismatch, so the mouth should be as large as practical to permit sound of wavelength less than its diameter to escape readily. The first job is achieved at the throat, and the second is achieved by the smooth way the horn enlarges its diameter, minimizing impedance mismatch, heading for a large opening.



If the horn receives sound instead of sending it, feeding a detector like a microphone or an ear in its throat, the principle of reciprocity is at work. The big buildup of amplitude in the throat will still apply.. Before radar, approaching enemy aircraft could be detected best with large horns aimed at the horizon and manned by humans, who no doubt had to hope that no sudden loud sounds would arrive.





Sound that won’t leave

The movie of the raindrop effect simulation can be downloaded here.


Initially a sound pulse started at the upper right, and lower frequency (longer wavelength) sound traveled to the bottom first, followed by higher frequency waves. This explains the chirp. The velocity of sound is not independent of wavelength for the sound trapped near the stairs. After a long time, the sound settles down to the pattern above, and in the movie, with orange and blue colors exchanging places every half period.

Computer synthesized sounds - in conjunction with computer animation.

Computer visual synthesis developed ahead of computer sound synthesis for the visual events being depicted, but sound is rapidly catching up.


SIGGRAPH is the annual computer interactivity, graphics, and sound conference - a gathering for industry (including fledgling start-ups), academic, and creative professionals - a good place in reality and online to see some cutting edge work.

Various shortcuts are used to render sound, ones that seem to be good enough, and have a physical basis. This is very much the case for visual animation as well. Certainly no one is starting at the molecular level to do either sound synthesis or animation! However even at the level of good science, equations for sound generation do not start there anyway, but rather have already lumped molecular behavior into a cellular picture as outlined in Chapter 1. The high road for sound synthesis would be to solve for the smallest pieces of solid, liquid, and gaseous material in a scene, from the smaller shards and droplets and the cells of larger objects coupling to the air and thence to the ear or microphone meters or more away. Alas, this is too difficult for computers except in very simple scenarios, thus the various "good enough" procedures.

Vocal tract computer voice synthesis; excellent web site by Peter Birkholz

More on the cookie cutter phenomenon:


In Norway, more such events quoted in this website:

"About 1 km SE of Skogvollvatnet (a lake), at Skogvollmyra (a moor), a slab of turf 5.2 m long and 1.8 m wide, has, in an apparently inexplicable manner, torn itself loose from its 'mother turf' and placed itself 4-5 m away. The slab of turf is completely undamaged and is placed with the right side up. The piece of turf has rotated 20-30 degrees compared to the original hole. The hole in the moor is absolutely even at the bottom, and the angle between the bottom and its walls is 90 degrees. The hole is 30-35 cm deep, and its edges are nicely cut.
"From the hole there is a crack running westwards for about 6 m. Close to the hole this crack is somewhat widened, and one side of the crack twists itself 25-30 cm above the other. This twisting decreases as one gets further from the hole. The crack gradually subsides, and it is hard to tell exactly where it ends.
"About 12 m NW of the hole there is an arched crack of about 15 m lying with its concave side towards the hole. It is plainest in the middle. Here the side closest to the hole has been twisted upwards about 15 cm. Here also the crack gradually disappears at both ends. There is an open hollow beneath the part that has been twisted upwards, about 30 cm below the surface.


Waves passing through random speed variations.

This is our culprit for the cookie cutter phenomenon, as described in Why You Hear What You Hear. The wave energy gets concentrated into branches inside of which there is much higher energy.

This is exactly what happened as the wave pulse from the 2011 Sendai tsunami spread out over the pacific ocean. The water surface wave speed variation is caused by the variable depth of the ocean. The long wavelength waves "feel" the bottom and travel slower over shallower water. This corresponds to a variable refractive index, and bending of the would-be uniform ripples into branches of flow. The result is shown below in a computer simulation which is quite accurate since the bottom depth is well mapped out and the origin of the wave is known precisely. Note the energy concentrated into branches. Note too the spur heading for the California coast. This spur actualy arrived and killed a person who had ventured out along a sand bar to watch the tusnami. It happened in Crescent City, California; waves of height 2.5 metres were seen there.

branched energy flow in the pacific due after the Japanese tsunami

Branched energy flow in the pacific due after the Japanese tsunami. Japan and the tsunami launch point at the upper left.


Project:Doppler effect

This video has a nice sound track with a clear doppler effect, analyzed in the sonogram with clear pitch changes. Use the sonogram to estmate the speed of the train, knowing that the speed of sound is 344 m/s. The sound file of the key part as the train passes is here.

doppler effect amtrack train

Doppler effect in Big Bang Theory: